1
Infrastructure – a sensible investment?
Contents
Page
Introduction and conclusions 2
Scope of this paper 3
History 3
Why institutions will consider infrastructure investment 4
The need for more infrastructure 5
Current infrastructure investments (statistics) 7
Risk-bearing investments by institutions 8
Regulatory requirements for institutions 10
Risks in infrastructure projects 11
Infrastructure bonds 15
Unlisted Infrastructure Funds 16
Direct ownership (freehold or leasehold) 17
Listed infrastructure equities 19
Overall portfolio construction 25
Practical issues 27
Possible future work 28
Further reading 28
Conclusion 29
Working Party members 29
2
Introduction and conclusions
1. This paper has been prepared by a working party which has been looking into the suitability
of infrastructure as an investment for financial institutions. The principal considerations which
potential investors will wish to take into account are discussed, including the risks involved and
the returns which can be expected. It is, however, only an introduction to a very complex
subject, and there are pointers to some further reading. The main conclusions we have reached
are as follows:
The world needs massive extra investment in infrastructure over the next 15 years. It
will be natural for institutions to participate in this investment, given their role and
function in society.
Investing institutions may have some opportunities to make relatively good returns if
they are willing to follow the methods of infrastructure investment which bear the
higher risks.
The risks could be significant and due diligence is needed.
However, there is currently a shortage of investable projects, with a resulting downward
pressure on rates of return.
Infrastructure investments offer institutions stable cash flows for many years,
sometimes with links to inflation. The capital values of the investments may be less
volatile than other categories of investment, which may help to make valuations of an
institution’s total portfolio more stable.
Investments in infrastructure bonds are not risk-free unless there are watertight
guarantees from credit-worthy governments.
The highest risks and potential returns are likely to come from direct ownership of
infrastructure assets (freehold or leasehold), or from the “equity” component of a
structured financing package for new infrastructure. However, the risks are sometimes
significant and hidden, and considerable due diligence is needed before authorising an
investment. The work involved is costly but can be shared out between institutions
which use common investment platforms. These platforms also have the advantage of
enabling institutions to invest in more projects, thereby getting a better distribution of
risk. The greatest risks usually arise when investments are made before construction
has commenced, and some institutions may prefer to wait until construction is finished
and the asset is in operation. Ownership may provide further development
opportunities to enhance returns. The position of the investor is somewhat similar to
that of investors in commercial property. Investors may need to work closely with
planners and other parties to find suitable investment opportunities.
Infrastructure funds have additional risks (such as those resulting from gearing) and
usually shorter timescales, and may therefore not meet the needs of many institutions.
The shares of listed utility and infrastructure companies offer a quick and easy route into
infrastructure exposure but possess too many of the characteristics of equities in
general to be seen as a significant diversification from those equities.
Regulatory requirements are currently a less important obstacle to institutional
investment in infrastructure than the shortage of suitable investments, but the
regulatory position needs to be clarified for the future.
3
Scope of this paper
2. By “infrastructure” we mean all the fixed assets which can be deployed for economic benefit,
whether they give rise to a revenue stream once operational or not. Many of these assets
require very substantial initial investment and are undertaken by public sector sponsors, but
some are undertaken by companies listed on the Stock Exchange, while others are financed by
private equity and private debt. For a large project the finance for the initial investment can be
raised in a variety of ways, including: bonds; inflation-linked bonds; development companies
backed by banks; specialist infrastructure funds financed by investors; sale of a freehold or
leasehold interest; and equity finance raised from investing institutions. Some projects are
financed from just one of these sources, but others have a structure in which tranches of the
investment are financed by investors who are willing to accept different degrees and types of
risk. Some of the areas of economic activity in which new infrastructure is often required
include transport, energy and power, communications, water, waste disposal, housing, and
commercial property such as shops, offices and warehouses. This last category is one with
which many institutional investors are already familiar, and some of them have large commercial
property portfolios, spread over a number of individual assets. Some infrastructure projects in
future may be proposed with the aim of making communities more able to withstand climate
changes, for example through flood protection schemes. In other cases the aims of public
sponsors of new infrastructure will often be to stimulate economic growth in a particular
geographical area. On the other hand, the aims of investors will usually be to make a return
which is sufficiently large to take account of the risks they have run. This paper discusses the
tensions which can sometimes arise between these sometimes conflicting aims. It also
considers the risks in some detail and the considerations investors might wish to think about
when trying to balance risk and reward.
History
3. Much of the infrastructure in the UK has been financed from private sources. For example,
in the 18th century many roads were impassable muddy tracks in winter and they were improved
by trustees who borrowed money to lay down hard surfaces and then repaid the debt from the
tolls charged to travellers. Georgian canal companies and Victorian railway companies financed
some massive infrastructure projects, in the hope of making a good profit once their borrowings
had been repaid. Docks, water and energy companies were also formed in the same way.
Many of these companies were nationalised in the 1940s. More recently we have seen the
privatisation of many utility companies, some of which have undertaken further big
infrastructure schemes. Other forms of infrastructure, for example hospitals and schools, have
been financed in the last 20 years through the Private Finance Initiative, with the aim of bringing
in efficient private sector management of the initial construction and the eventual ongoing
operation of the facilities. The Government has sometimes had to step in to encourage private
investment to come forward, an example of which is provided by the multi-billion investment
programme being undertaken by Network Rail since 2001, which is financed by bonds with a
Government guarantee. In 2012 the Government set up a scheme to enable up to £40 billion of
4
other infrastructure to be financed by bonds with a Government guarantee, but so far projects
worth only about £4 billion have been approved (reported comment by Treasury spokesman,
April 2015). It is planned that £17 billion of the sum available under the scheme will be used to
finance the rebuilding of the Hinkley Point C nuclear power station, but in April 2015 it was
reported that there are now serious doubts about the quality of the steel which will be used to
make a casing round the reactor. One of the advantages claimed for government guarantees is
that they enable funds for public projects to be raised more cheaply than in other ways.
Naturally the benefit of this will depend on the extent to which the market believes the
guarantee can be relied upon for many years into the future, which will vary from one country to
another.
Why institutions will consider infrastructure investment
4. Some of the reasons why financial institutions might wish to invest in infrastructure bonds or
equity as part of their overall portfolio are:
a. A possibly better return than on other investments;
b. Stable and predictable cash flows during the operations phase;
c. Returns on investments which are insensitive to fluctuations in business, interest rates
and stock markets and have low correlations with the returns on other investments;
d. Need to diversify in order to get a more stable return on the portfolio as a whole;
e. Need to match long-term liabilities with long-term income streams;
f. Requirement for a return which is inflation-linked over a long period (where applicable);
g. Relatively high recovery rates, low default rates and good credit ratings; and
h. Desire to enhance reputation by being seen to finance social infrastructure.
The hurdles include:
Regulatory requirements for investors;
Illiquidity of many infrastructure assets;
Lack of transparency;
Need for lengthy and costly due diligence;
Poor availability of statistics of past performance;
Fee levels for intermediaries;
Lack of understanding of risks;
Uncertainties which cannot be foreseen; and
Difficulty of getting to the front of the queue for the best investments.
Potential investors will need to consider whether the returns on investment will be lower
than has been the case historically and the extent to which investors are being rewarded for
credit risk as well as liquidity risk. They will also need to consider whether the risk-reward
balance on infrastructure is sufficiently attractive, in comparison with the risk-reward
balance on other kinds of investment.
5
The need for more infrastructure
5. As economies grow, more infrastructure is always needed. How much investment is likely
to be required globally? Is a funding gap likely? The expectation is that increased
investment from insurance companies and pension funds may be needed, to fill the gaps
which governments leave unfinanced. It is possible that, unless financial institutions in
developed countries increase their readiness to invest in infrastructure, some of this money
will not be found and world economic growth will be impeded. Perhaps the role of
governments round the world will change, so that instead of investing in infrastructure
directly themselves, they provide guarantees which will unlock greater flows of cash from
institutional investors.
One of the underlying considerations is that substantial infrastructure will be required to
transition to a low carbon economy and to assist with adaptation and resilience to climate
change. Whilst the directly attributable effects of climate change are likely to be
experienced in different regions in different time periods, many effects are likely to be
experienced well before 30 years’ time. The nature of infrastructure , in terms of the
investment of financial capital, natural resources, time and human capital, means that the
choices made now will have a substantial impact on whether and if so, when a transition to
a low carbon, sustainable development path is followed.
It has been estimated that the world will require public infrastructure spending totalling
US$ 2.8 trillion per annum, i.e. about 3% of world GDP, over the 25 years between 2005 and
2030 [see OECD study Global Infrastructure Needs to 2030, page 54;
http://www.keepeek.com/Digital-Asset-Management/oecd/economics/strategic-transport-
infrastructure-needs-to-2030/global-infrastructure-needs-to-2030_9789264168626-4-
en#page5]. About one quarter of this requirement is related to the distribution of water.
A more recent estimate comes in a PWC report [Capital Project and Infrastructure Spending
– Outlook to 2025, based on research by Oxford Economics, apparently published in 2013.]
This states that worldwide infrastructure spending will increase from $4 trillion in 2012 to
more than $9 trillion per year by 2025, the growth being mainly accounted for by increased
spending in China, Asia and emerging megacities. The differences between this figure and
the OECD figure in the previous paragraph may be partly accounted for by different
definitions of “infrastructure”, though this is unclear.
It is certainly theoretically possible that insurance companies and pension funds in OECD
countries might be able to contribute a sizeable proportion of the total infrastructure
investment required each year, since their assets totalled US$ 44.5 trillion in 2011 [Source:
OECD Working Paper No.36, Institutional Investors and Infrastructure Financing, 2013, page
8], but the question is whether they will decide to do so, given the risks and returns
available.
The European Commission estimated in 2013 [http://europa.eu/rapid/press-
release_MEMO-13-611_en.htm] that 1500 to 2000 billion euros would be required to
finance infrastructure project needs in Europe up to 2020. This equates to about 250 billion
euros per annum. They stated that it was difficult to assess how big the investment market
6
was at the moment, because it was so fragmented, but it was possible to make some
estimates - infrastructure transaction volumes had been stable at between 100 and 150
billion euros each year since 2007, indicating a consistent demand for financing.
In the UK there is a widespread recognition that more and better infrastructure is highly
desirable, to replace worn-out structures, meet new emissions targets, provide resilience in
the face of uncertainties, and satisfy the numerous needs of a growing population which
also contains more children and older people. There is a feeling that changes are required
in order to identify needs and suitable projects more systematically, and to achieve results
within acceptable costs and timescales. The Armitt review is proposing an independent
National Infrastructure Commission (NIC) to look 25-30 years ahead at the UK’s
infrastructure needs and to recommend a clear pipeline of projects with priorities (for
example, flood protection and energy supply). Following a Parliamentary vote on the
priorities, the Government would formulate detailed 10-year sector plans, and the NIC
would report annually on their delivery. These arrangements would replace the current
system where Infrastructure UK sits as a unit within HM Treasury.
Over 60% of world infrastructure investment in 2011 was attributed to Western Europe and
the USA, although this is expected to decrease as emerging rapid growth markets take a
larger share: as a result of higher expected growth rates, these markets’ relative share in
global infrastructure spending is expected to increase from current levels, financed at least
partly by an increased rate of insurance and pension savings by the populations of those
countries. One key question is whether local sources of finance will suffice to meet
infrastructure needs in those countries or whether contributions will also be needed from
financial institutions situated in countries which are already well developed.
Developing countries now spend about US$ 1 trillion a year on infrastructure, but
maintaining current growth rates and meeting future demands would require investment of
at least an estimated additional US$ 1 trillion a year through to 2020. [World Bank Press
Release, 9 October 2014]. The World Bank has emphasised the need not just to increase
the quantity of infrastructure but also to focus on its quality. Private infrastructure
investment in emerging markets and developing economies dropped from US$ 186 billion in
2012 to US$ 150 billion in 2013. The Bank said that the real challenge is not a matter of
money but a lack of bankable projects. In addition, of course, there is a continued need for
infrastructure investment in economies which are already developed, though here too there
is currently a shortage of investable projects However, many of the potential investors are
not yet ready to take the plunge, even if sufficient numbers of suitable projects were to
become available. At present the majority of institutions’ infrastructure investments consist
of low-risk and very low yielding bonds which have either a Government guarantee or a high
credit-rating, and it is unclear whether institutions will wish to venture out into other forms
of infrastructure finance in order to increase yields, where this involves extra risk.
7
Current infrastructure investments (statistics)
6. Statistics are available of the infrastructure currently held by investors:
Investors with assets under management of over $1 billion. OECD investors held USD 70
trillion in 2013 in long-term investments [Long-term financing of the European economy,
European Commission, 2013] backing assets with a life spanning 25 to 60 years, generally in
infrastructure. Institutional investors currently invest an average of roughly 3% of their total
assets in infrastructure financing. Of 170 Europe-based private-sector pension funds
investing in the infrastructure asset class, the top 25 funds, by allocation to infrastructure,
have aggregate assets under management of more than €392bn. On average, these top 25
private sector pension funds have a current allocation to the asset class of 3.5%, lower than
their average target allocation of 5%. With regard to how these investors allocate to
infrastructure, 14 of the top 25 Europe-based private sector pension funds have established
separate infrastructure allocations, while two invest through a general alternatives
allocation. The remaining six investors invest in the asset class through part of a real assets
or private equity allocation, part of an opportunistic investments allocation, or other types
of allocation. For future investments, the funds will consider different routes to market,
although all 25 of the top private sector pension funds favour investment via unlisted funds.
21% will consider investing in infrastructure assets directly, and 8% will do so via listed
funds. Direct investors in the infrastructure asset class have an average of $66bn in total
assets under management [Prequin, Jan. 2015].
Investors with assets under management of less than $1 billion. 301 smaller institutional
investors, with AUM under US$1bn, have aggregate assets under management of US$
125bn, and allocate an average 5% of total assets to infrastructure, which is below their
target allocation of 7%. These investors’ allocations are distributed between four main
types: part of real estate allocation (43%), part of private equity allocation (25%), separate
infrastructure allocation (19%), part of general alternatives allocation (8%), other (4%). Of
the separate infrastructure allocation, unlisted funds are the favoured way to access the
market, and only 7% of these smaller investors utilize direct investments.
Infrastructure funds. Entering Q1 2015, 454 infrastructure fund managers globally are on
the road looking to raise an aggregate target of US$ 102bn from institutional investors
[Prequin]. This includes both direct and indirect primary funds offering both debt and equity
participation. In 2014, US firms maintained their lead in terms of capital formation, raising a
total of $50.44 billion, with Australia following closely behind with $47.32 billion. Europe is a
distant third, accounting for $21.32 billion – with France the leading country accounting for
almost $10 billion – followed by Asia with a total of $9.58 billion. Though these figures are
impressive, it is clear that infrastructure funds currently contribute a relatively small
proportion of global infrastructure spending.
8
Risk-bearing investments by institutions
7. This section gives details of some of the known risk-bearing infrastructure activity by
financial institutions but does not purport to be fully comprehensive. It excludes investment in
commercial property, which is a form of infrastructure investment that many pension funds and
insurance companies have undertaken for many years on a risk-bearing basis.
Some of the ways in which institutions have already funded infrastructure round the world are as
follows:
i. Many Governments issue bonds to finance infrastructure construction. Where these bonds
have a solid Government guarantee, the infrastructure connection is purely incidental and they are
treated by institutions in much the same way as ordinary Government bonds in which they
habitually invest. Sometimes projects are financed by a private loan arrangement with an
institution.
ii. Australian and Canadian pension funds have been among the most active investors in
infrastructure, both at home and overseas, with some having allocated 10% or more of their
investment portfolio to infrastructure on an “equity” or risk-bearing basis. They have acquired
the necessary knowledge and expertise and have in-house resources able to carry out research
and risk-assessment of projects. They can co-invest with other parties, often from the public
sector in a “public-private partnership” and they sometimes even take leading roles. They have
been particularly active in energy assets.
iii. Until recently UK pension funds and insurance companies have not usually invested in
infrastructure equity or ownership to any great extent, although one or two of the largest pension
funds have dipped a toe into equity infrastructure opportunities – for example the Universities
Superannuation Scheme bought Australia’s Brisbane Air Train Service for $110m in 2013.
Some pension funds may have small investments in specialist infrastructure funds, which increase
the risks and returns by gearing and typically have a 10-year time horizon and large fees. There
have been two important developments in the last 3 years:
A new Pensions Infrastructure Platform (PIP) was established in 2012 by the National
Association of Pension Funds to attract low-risk investment by pension funds in new
roads, hospitals, airports, etc. The PIP will favour mature “brownfield” investments
with a detailed history of use, including existing public-private partnerships, and is
unlikely to consider new “greenfield” investments unless construction risk can be
reduced. In November 2014 it was reported that the platform had so far committed
close to £300 million, with a target of £2 billion.
In 2013 insurance giants including Legal & General, Prudential, Aviva and Standard
Life pledged £25 billion of UK infrastructure investment over a 5-year period as part
of the government’s re-launched National Infrastructure Plan. It is known that Legal
and General, for example, have invested in student accommodation, and have also
been supportive of small projects with big local impacts. Aviva’s Investment
Management arm has invested in domestic solar panels. However, there has been a
shortage of long-dated investments which will match an insurance company’s
liabilities. Rothesay Life announced in April 2015 that they were likely to cut back
on planned investment in 2015 because not enough infrastructure investments had
come forward under the UK Government Guarantee scheme (see paragraph 3).
9
iv. The infrastructure investment body Scottish Futures Trust has introduced the Hub model for
smaller-scale public-private community infrastructure projects, under which public sector bodies
form ‘hub companies’ with the private sector, as procurement vehicles for projects. A project
can include contributions from multiple public bodies – for example, local authorities, health
trusts, police and fire authorities – which will share use of the resulting facility. [PMI News, April
2015, page 37].
v. Pensions Insurance Corporation UK is a low-risk investor with £12 billion of assets. It is
seeking long-dated infrastructure bonds with investment grade credits to match its liabilities,
including bonds with inflation protection.
vi. OECD reports have concluded that European pension funds have only recently started
building up their investments in infrastructure, with allocations ranging up to about 3% of their
total portfolio. They usually prefer to invest in large, mature operating assets that already
generate cash flow, without taking construction risks. However, there are signs that some of the
largest European pension funds are starting to form joint platforms for infrastructure investment,
somewhat along the lines of the UK’s PIP, and in particular are taking an interest in “green”
projects.
vii. The National Pensions Reserve Fund of Ireland is seeking to leverage its 7 billion euros of
assets to achieve Irish investments worth 21 billion, aimed towards stimulating national economic
growth. The Fund is willing to come in at the start of the asset creation process, with higher risk.
It is, however, willing to partner with pension funds requiring lower risks and will create
appropriate structures to facilitate this.
viii. GIC Special Investments Pte of Singapore invest in the “equity” of infrastructure, as an
arms- length manager for the Government of Singapore, anywhere in the world outside
Singapore. Employing a team of about 20 people, they are willing to take a degree of risk and
are seeking the best possible risk-adjusted returns.
ix. Public Private Partnership models have long been used in the UK for projects such as new
hospitals and schools, where a private sector consortium constructs, equips and manages the
building, in return for a rental paid by a local health trust or education authority which depends on
availability of the facility and the achievement of quality standards. At the end of a
predetermined period, such as 30 years, the ownership of the building reverts to the public sector.
The extent to which such developments have been financed by institutions is unclear. In some
cases the private investors may be able to identify opportunities to make extra returns, for
example by letting out schools in the evenings or by designing in additional facilities such as
shops. In some cases the deals have been criticised as giving too high a return to the private
investors, so a new model has been developed by the Scottish Futures Trust for projects worth
over £50m. This is known as the Non-Profit Distributing (NPD) model, which caps the returns
available to investors [PMI News, April 2015, page 37], and apparently six deals have been closed
so far. The model is being considered for the proposed £200 million Velindre specialist cancer
centre in South Wales. In view of the cap on returns, it seems likely that such deals will only
interest insurance companies and pension funds for tranches of finance where the risks are
perceived as minimal.
x. Some investment managers offer investors a portfolio of listed infrastructure shares, e.g.
water and energy companies.
10
xi. Banks have also been investors in infrastructure. Some reports suggest that this activity may
be tending to decrease, because of new solvency requirements. Other reports say that bank
lending for project finance increased in 2014, and that institutional investors are finding it
challenging to match the speed and deal-flexibility which banks offer.
Regulatory requirements for institutions
8. At present regulation issues are less of a deterrent to infrastructure investment by insurance
companies than the lack of supply of suitable infrastructure investments, which is pushing up
prices and compressing yields. The regulations themselves are only really a strong barrier in the
sense that, if the assets are being used for the Matching Adjustment to back annuities, then this
may restrict the volume of eligible assets, just as it is an issue for other fixed income assets like
corporate bonds etc. For non-Matching-Adjustment assets, although there is a feeling that capital
charges may be higher than they should be, this is not stopping insurers from investing when they
find assets carrying suitable risks and returns. However, the regulatory issues are to some extent
creating a gap between the supply of infrastructure investments in the market and the particular
requirements of insurance companies. Some insurers want long term infrastructure debt to match
liabilities but may also require Spens clauses to obtain matching adjustment eligibility. This
locks the infrastructure equity holder into a long term commitment, with arguably penal Spens
clauses in the event that refinancing is required. [A Spens clause applies to UK listed bonds and
provides protection to the investor, by ensuring that on an early termination of the bond the
investor receives sufficient compensation to enable it to obtain the same cash-flows by re-
investing in risk-free gilts.]
9. However, in the longer term there is a case for considering whether there should be special
treatment for infrastructure investments in the regulatory framework. EIOPA (the European
Insurance and Occupational Pensions Authority) has commenced a study:
to develop a definition of infrastructure investments that offer predictable long-term
cash-flows and whose risks can be properly identified, managed and monitored by
insurers;
to explore possible criteria for the new class of long-term high quality infrastructure
assets covering issues such as standardisation and transparency; and
to analyse the prudentially sound treatment of the identified investments within
Solvency II, focusing on their specific risk profile.
[Source: EIOPA newsletter, February 2015].
10. EIOPA’s initiative is consistent with the aims set out in a Green Paper, Building a Capital
Markets Union, published by the European Commission on 18 February 2015, which states that
the challenge is to unlock investment in Europe’s companies and infrastructure.
[http://ec.europa.eu/finance/consultations/2015/capital-markets-union/index_en.htm]. The stated need is to make the system for channelling investment funds – the investment chain – as efficient as possible, both nationally and across borders. The recently finalised European Long-
Term Investment Funds (ELTIFs) regulatory framework may also help to make long-term
investment in infrastructure easier. The European Commission believes that ELTIFs should have
particular appeal to investors such as insurance companies or pension funds which need steady
income streams or long term capital growth.
11
11. EIOPA launched a consultation on 27 March 2015 through their Discussion Paper on
Infrastructure Investments by Insurers and are expected to publish a formal response to the
European Commission by the end of June 2015. The paper asks what elements in the Solvency II
framework might prevent insurers from investing in infrastructure, and discusses how these
elements could be adapted while preserving the same level of policyholder protection. The paper
points out that investments in infrastructure present potentially complex risks, which can also vary
significantly between different types of infrastructure projects, and that there should be sufficient
understanding of the assumptions upon which future performance or usage of the project, and thus
payments to the financiers, are based. It also states that not only is a thorough due diligence
process vital before deciding upon an investment, but such investments will also require active
engagement to monitor their ongoing performance and suitability. The paper discusses how
insurers should satisfy supervisory bodies that they are dealing adequately with such issues,
perhaps by using an internal rating system. Is there scope for insurers to work together co-
operatively in developing rating systems which will satisfy regulatory requirements? Will
Matching Adjustment and hold-to-maturity restrictions limit future investment in infrastructure by
insurers? Will social and political factors be given any weight in the decision as to whether a
particular infrastructure investment is acceptable?
Risks in infrastructure projects
12. This section looks at the risks for investors in infrastructure projects. These risks will be
considered under 6 main headings:
Construction risks
Forecasting risks
Operating and revenue risks
Guarantor credit risks
Financing risks
Other risks
Each heading will now be considered in turn, and the section concludes with some remarks on the
risk consequences of illiquidity, the possibilities for risk mitigation, and the need for risk
investigations and risk governance.
13. Construction risks. These are the risks which an investor bears if he takes ownership of land
and constructs a physical infrastructure asset on it. To start with, there is the risk that planning
permission may not be granted. Even once this has been obtained, the cost of the project may be
hard to estimate. Because of the uncertainty of the issues which may arise during the
construction process, it is sometimes difficult to forecast, even approximately, how much the
eventual cost of construction will turn out to be. For example, the Edinburgh trams project was
originally costed at £375m in 2003, but the budget was later increased to £545m. By June 2011
it was apparent that the project had gone seriously wrong and it was decided to cut out the eastern
section of the line, but even so the eventual cost was over £770m for the reduced project. In
many projects, however, construction costs come in quite close to the estimates made after the
project plan has been completed but before construction commences. For example, two current
projects – Crossrail and the Second Forth Crossing - are looking good in this respect, though in
neither case is construction yet finished. There are also potentially expensive risks which may
arise from social or environmental causes during the construction process, including deterioration
of water quality over the surrounding area, and the possibility of delay and extra cost due to
12
opposition from special interest groups concerned with noise, landscape protection, preservation
of historic buildings, etc.
14. Forecasting risks. If the investment is being made before the project comes into operation,
there is always the possibility that some of the key forecasts which have been made about the
period after construction is completed may turn out to be wide of the mark. These forecasts
typically relate to usage of the new infrastructure, and the ongoing net revenues from it, after
deducting operating, maintenance and renewal costs. These net revenues might turn out to have
been very much over-estimated, either by accident, by the difficulty of predicting the future, or
sometimes as the result of deliberate bias by those anxious to see the project proceed. As an
example, at the time the decision was taken to build the Channel Tunnel (around 1985), 15.9
million passengers were predicted for the Eurostar trains in the opening year (1995), but the actual
numbers were only about 2.9 million, rising to 7.1 million in 2000 and falling to 6.3 million in
2003. The completion of a high-speed connection on the British side (the Channel Tunnel Rail
Link, from London) then caused Eurostar traffic to rise, but by 2014 there were still only 10.4
million passengers.
15. Operating and revenue risks. Once the infrastructure asset has been completed and a stream
of net revenue has been established, an institution looking for a low-risk investment might be
willing to purchase the revenue stream, perhaps on the basis of taking a 30-year lease. There will
still be some remaining risks, however, and the investor will need to consider these very carefully.
First, the gross revenues, if they derive from fares or charges based on actual usage, might be
adversely affected in future in an economic recession or if competing infrastructure is built which
results in reduced usage. Even if there is some comfort from a regulatory regime that guarantees
fixed prices for a number of years (for example, for electricity generated by a wind park), there
may be uncertainty about the revenue stream thereafter. For any kind of infrastructure, operating
and maintenance costs might rise due to wage increases or new safety requirements introduced by
regulators. It is possible that the revenue stream may be temporarily interrupted by severe
weather, fire, earthquake or terrorism. Essential equipment might have to be replaced earlier than
anticipated if it becomes prematurely worn out or obsolete. Faults may be discovered in the
construction of the infrastructure, which require costly remedial action and enable only limited
recourse (if any) against the contractors who built it. The EIOPA discussion paper (see paragraph
11) has an annex which discusses possible criteria for assessing revenue risk for the purpose of
insurance regulations.
16. Guarantor credit risks. If the return to the investor is underwritten by a guarantee from a
credit-worthy government or other third party, for example where an infrastructure bond has been
issued, the operating and revenue risks referred to in paragraph 15 can be largely discounted,
since the cost will fall on the guarantor. In some cases, however, the guarantee might have been
issued by a subsidiary company which does not have automatic backing from a credit-worthy
party and in such cases it may appear that little value can be placed on the guarantee.
17. Financing risks. The capital structure of some “private equity” funds which aim to own
infrastructure may involve an element of gearing, where the fund borrows money in order to
increase its infrastructure exposure and increase the expected return to fund participants. One
risk here is that the borrowing might have to be rolled over during the fund’s lifetime at a higher
rate of interest, leading to a diminution in the investors’ returns. Another risk for these
infrastructure funds, which often have a life of 10 years or less, is that when the assets come to be
sold at the time the fund is wound up, the prices achieved may be very much less than expected,
13
so that investors make a loss. If the prices achieved are very low, there might even be the risk
that investors will have to put more money into the fund at that time in order to repay all the
borrowings. Gearing may also exist in the capital structure of the financing arrangements for an
infrastructure project, where the major part of the cost has been financed by issuing bonds and the
remainder by issuing equity-type instruments, and the holders of the latter may experience very
variable returns (in either direction).
18. Other risks.
It cannot be sufficiently stressed that a major infrastructure project is often propelled
forward by the intense efforts of a politician or senior official whose career depends on
the project getting approval to go ahead. This may well lead to pressure to get the
project approved before the costs have been fully evaluated and the risks properly
assessed; moreover, the project may not be the optimum way of achieving the benefits
envisaged because it has been selected too hastily. In some cases there may be hidden
motives, such as seeking work for construction companies and creating jobs for workers.
These factors could well lead to bias in the cost estimates for the project or in the
estimated usage or ongoing costs, which will only become apparent at a later stage. Even
after project approval there is a risk that there may be a change in the political situation of
the country where the infrastructure exists, leading to the abandonment or major
modification of the project while under construction or to new requirements being placed
on the operators of existing infrastructure (for example the imposition of stricter rent
control on residential accommodation or control of the fares which may be charged in a
transport project).
In the case of investment in infrastructure overseas, there are the risks of currency
fluctuations, repatriation of money and sanctions. When the infrastructure reaches the
end of its life, there is the risk that removal and clean-up may cost more than anticipated.
There is sometimes a danger that an infrastructure project will give rise to significant
hostile acts by parties who are adversely affected by it – an example would be the
diversion of a river by a country through which it flows upstream, to the detriment of a
country through which it flows downstream. There are also risks from terrorism or cyber
attacks for other reasons.
Given that many infrastructure investments will last for 30 years or more, it is necessary
to consider the implications of likely climate changes, many of the effects of which are
likely to be experienced long before the expiry of this period. These may well include an
increased frequency of severe weather events, with consequences such as flooding, and
the extent of the increase may be hard to quantify with confidence. In some cases higher
temperatures may adversely affect usage, if the facility is not sufficiently adaptable,
although in other locations the changes might actually be beneficial. In future it will be
desirable for infrastructure to have a degree of flexibility and resilience built in, even at
extra cost, so as to reduce the chance of costly repairs or even premature obsolescence
due to climate change, as well as enabling it to recover quickly from any disruption
caused by hostile interests.
19. The risk consequences of illiquidity. Many infrastructure investments are illiquid, in that it is
difficult or impossible for the investor to sell their interest before the expiry of the period of time
for which the contract runs. This may cause risks for the investor, if its own circumstances
change before that point is reached and cash is needed prematurely, and (unlike many other kinds
14
of investment) it also has the effect of making it hard to walk away if any of the risks listed above
show increasing signs of materialising.
20. Risk mitigation. If the investor is undertaking responsibility for the construction of the asset,
as well as its eventual operation, it may be worth insisting on the use of a methodology such as
RAMP (Risk Analysis and Management for Projects) to resolve uncertainties, help in the design
of the asset, keep risks to a reasonable minimum and determine the cost-effectiveness of various
risk-mitigation actions. If the contractor who is responsible for building the asset is willing to
enter into a fixed-price contract, this will go a long way to reducing the construction risk for the
investor, although even in a fixed-price contract there may be extra costs for the investor to bear if
specification changes have to be made. It will be important to check that the contractor will be
able to finance any cost over-runs himself without any risk of going bankrupt, since if the latter
occurs and a new contractor has to be engaged, this will inevitably result in considerable delay
and extra cost for the investor. Forecasts of net revenue streams can be checked to some extent
against the outcome for similar projects elsewhere. An investor can avoid construction and
forecasting risks by adopting a policy that no investment will be made until the asset is complete
and a revenue stream has been established, even though this will reduce the expected return.
Because of the somewhat unpredictable nature of future revenue streams depending on usage,
some investors might prefer to invest in projects where a fixed revenue stream will be payable by
a credit-worthy public body just because the asset is available for use: in such cases the main
risks will relate to maintenance and renewals of equipment. Many of the other kinds of risk
require a process of due diligence, differing from one investment to another, to tease out the exact
nature of the risk and make some broad estimates of likelihood and impact. For some risks it
may be possible to insist that they are borne by other parties or even insured (as in the case of fire,
for example). Finally the documentation must be examined very carefully indeed to make sure
that the investor is not inadvertently assuming unanticipated risks, which could cause the
investment to have a zero or even negative value.
21. Investigation of risks. It will be apparent from the above discussion that the risks for
investors in infrastructure projects can vary greatly from one situation to another. The highest
risks will usually be borne by investors who come in at the very start of a project and are prepared
to bear construction and forecasting risks, as well as the risks of operation and maintenance once
construction is complete. These situations will also normally offer the highest potential returns,
and sometimes these returns will be able to be enhanced by exploiting opportunities for additional
development associated with the main infrastructure (for example, shopping centres associated
with transport infrastructure). Potential investors in these situations will need to have a deep
understanding of all the risks to which they will be exposed, though to some extent the risks can
be spread by investing in a number of different projects. Concentration risk tends to be higher
for infrastructure than for traditional equity or debt because of high unit sizes. For lower-risk
investment opportunities, probably offering lower returns, whether of a bond or equity nature,
there may still be hidden risks which need to be understood before investment is committed. The
investigation of risks will always be a costly process, if it is done properly. One option for an
investor who does not have his own specialised knowledge is to go into partnership with another
investor who is committing his own money and will carry out a risk investigation on behalf of
both parties. Platforms are being constructed which will enable like-minded pension fund
investors to share the costs of risk investigations between them and to share in a number of
different long-term investments in order to spread the risks.
15
22. Risk Governance. The risk governance surrounding infrastructure investment should cover
the full lifecycle of the investment, from underwriting through to monitoring as well as possible
impairment. Ultimately any qualitative investigation of risks may need to be converted into an
internal rating in order to facilitate the selection of investments, the monitoring of risks and the
setting of capital requirements. Investors are likely to need a risk investigation framework and
internal rating system to ensure that projects are investigated and monitored consistently, perhaps
using Key Performance Indicators and Key Risk Indicators. The heterogeneous nature of
infrastructure investments, both in terms of the underlying project and the legal and financial
structure, add to the challenge of comparing, rating and monitoring different infrastructure
projects. Consideration will have to be given to what the investor would do to control any risks
which materialised, including any refinancing options. What will the insurer do if an investment
becomes impaired? For insurers with an Internal Model, infrastructure will need to be
incorporated and this will require risk calibration; there will also need to be a feedback loop
between the risk monitoring activities and the risk model, and new techniques may be needed to
include infrastructure in models of Economic Capital.
Infrastructure bonds
23. Infrastructure bonds are bonds issued by various bodies to finance the construction of new
infrastructure and because there is usually an element of risk, they normally offer a higher yield
than government bonds. The key question for investors will always be whether the yield is
sufficiently high to compensate for the risks involved, including any risks relating to the credit
quality of the issuer.
An example of what can go wrong has been reported. The Castor project was a natural gas
storage facility off the east coast of Spain. Work was financed by a 1.4 billion euros issue of 30-
year 5.8% bonds by a consortium of European banks. After gas was first injected, earthquakes
were detected in the area and work on the project was halted in 2013. The project was later
abandoned. In this particular case the bond holders got their money back [Catalan News Agency
3 October 2014] because they were bailed out by the Spanish taxpayer, but it is easy to see that
this might not have occurred had the contractual arrangements been worded differently.
It is important to note that the credit quality of the issuer needs to be investigated fully, and in
particular whether the issuer may be a subsidiary body, the solvency of which is not fully
guaranteed by a credit-worthy parent body.
Atlanta is a US city that has voted, in March 2015, to issue $250 million of 30-year bonds to start
to cope with a backlog of infrastructure projects costing four times that figure. About two-thirds
of the money will be spent on transportation schemes, street resurfacing, etc. Much of the sum
raised will be spent on projects of a city-wide nature but there will also be a large number of
purely local schemes. Property tax revenues will be pledged as security for servicing the bonds.
The bonds have not yet been issued but in January 2014 it was projected that the interest rate
would be 5.4% per annum [Renew Atlanta 2015 Infrastructure Bond web-site, FAQs]. The
return to investors, if this is the interest rate decided upon and all goes well, will clearly be
significantly higher than the 3% per annum currently obtainable for an investment in US
government 30-year bonds. No doubt potential investors, or ratings agencies on their behalf, will
be doing due diligence on the prospects for the city’s property tax revenues over the next 30
16
years, taking into account the likelihood that the infrastructure projects themselves may improve
those prospects. They will want to investigate also whether there will be any prior or pari passu
claims on those tax revenues for other purposes.
The municipal bond market in the UK is still in its infancy, but the first bond issue on behalf of
the Municipal Bond Agency, covering a number of local authorities, is due to come to market in
the next few months [Professional Pensions, 26 March 2015, page 27]. The motivation is to
reduce the cost of borrowed money below the current margin of 100 basis points over gilts
charged by the Public Works Loans Board, where local authorities borrow 75% of their money at
present. It is not yet known how far the proceeds of the bonds will be used to finance new local
infrastructure projects, and how much of the money raised will be used for other purposes.
Whether the return on the bonds will be sufficient to attract institutional investors, given the risks,
is currently uncertain.
Where infrastructure bonds have a watertight guarantee by central government, the considerations
for investors are likely to be very much the same as for bonds actually issued by that government.
There may be a slight improvement in the rate of return, if there are liquidity issues.
Unlisted Infrastructure Funds
24. One of the practical ways in which institutions can invest in infrastructure is through
unlisted infrastructure funds, which offer the prospect of a significantly higher rate of return than
infrastructure bonds, with different risk and duration characteristics. Although fees are high, the
due diligence on individual assets is undertaken by the fund’s management and the investor does
not need its own dedicated resource specialising in infrastructure. Many of the funds last for only
about 10 years, the intention being to sell the assets on to other parties in due course, so this is not
a long-term investment. There is usually a lack of liquidity before the fund is wound up. In one
way the risk is reduced, since the funds typically have a number of assets in their portfolio, which
offers diversification of risk. However, the risk is increased in other ways because of the
significant factors referred to in paragraph 17, as well as the currency risks if some of the assets
are located overseas.
The choice of suitable funds in which to invest is a non-trivial task, since it is necessary to try to
get a clear picture of the nature and extent of the risks which the fund will be running, and in
particular any construction or forecasting risks, or “key-man” risks in the fund’s management
team. An important issue will be whether the aims of the management team are aligned with
those of the investors because the managers have their own money in the business.
17
Direct ownership (freehold or leasehold)
25. Direct ownership of an infrastructure asset, either freehold or leasehold, is in some respects
similar to owning a commercial property. There are, however, some extra risks and in addition
the opportunity to sell a commercial property does not usually exist in the case of infrastructure.
These direct unleveraged investments in infrastructure arguably have some of the characteristics
of debt in terms of the stability of cash flows and some of the characteristics of equity in terms of
open-ended risks and upside potential. The capital treatment of a direct investment is unclear
under the Standard Formula but may be no better than an equity investment, creating an incentive
to restructure the asset.
26. The first case to consider is where the investor is willing to buy the land on which an
infrastructure asset is to be constructed. He might be willing to bear the cost of construction
himself, perhaps engaging a contractor on a fixed-price contract. Once construction is complete
he will then be responsible for the operation and maintenance of the infrastructure, probably
offering contracts for these tasks to specialist operators and maintenance companies, and he will
get a revenue stream depending on usage and charges. This type of investment is where the risks
and the potential returns are likely to be highest.
27. If the investor does not currently own the land, and wishes to commit to the opportunity but
avoid construction and forecasting risks, he might be willing to give an advance commitment that
at some future date, when the asset is finished and a short period of successful operation has been
completed, a long lease of the asset will be purchased at a predetermined multiple of the net
revenue stream that has been established. There is a precedent for this in the town centre
shopping malls constructed in the UK in the 1970s, where pension funds would guarantee before
construction started that when the asset was 85% let they would buy out the investment on a long
lease at a predetermined multiple of the rents achieved. An example where transport
infrastructure was purchased by investors after the asset was complete is provided by the Channel
Tunnel Rail Link, where two Canadian Pension Funds bought it on a 30-year lease for £2.1 billion
in 2010, contracting track management, maintenance and renewal to Network Rail (though in this
case there was no pre-commitment).
28. It may be that the investor would prefer to invest in property associated with the new
infrastructure, rather than in the infrastructure itself. For example, in the case of a new transport
link, there will often be opportunities to build shopping centres at principal stations. These
investments will share many of the characteristics of normal commercial property investment,
though there may be additional risks associated with usage of the infrastructure on which they are
centred.
29. As an example, a large Malaysian pension fund provided the capital to buy London’s disused
Battersea Power Station, which will be developed by two commercial companies in partnership
with the pension fund., This consortium is making an up-front capital payment to finance part of
the cost of constructing the Northern Line Extension, a new underground railway which will serve
the developed properties on the Power Station site and in the surrounding area. It is expected that
the new railway will stimulate the development of other properties in the surrounding area, and
pension funds will have the opportunity, alongside developers and other investors, to participate
in those developments through the normal commercial property market.
18
30. Other types of freehold or long leasehold investment which are available include energy
installations, residential property (buy to let), student accommodation, and private finance
initiatives (hospitals, schools, prisons, etc). Each of these will have its own risk characteristics,
and much will depend on whether the investor has to bear construction and forecasting risks. In
many cases the return will depend on actual net receipts from the infrastructure once it is in
operation, but for some types of infrastructure it may not be practicable to generate a flow of
income and in these cases a shadow toll depending on usage or availability may be payable by a
government body. These shadow tolls may have an inflation linkage built in, which could be
attractive to long-term investors who have inflation-linked liabilities.
31. Investment into the energy sector, and notably renewable energy assets, has seen interest by
institutional investors. Examples include investments through project loans, bonds or equity
participation. For example, an onshore windfarm was commissioned in Jädraås, Sweden in
2013, at a capital cost of €360m. Of this, €120m is provided by Pension Denmark in the form of
15-year synthetic loans with a guarantee by the Kingdom of Denmark. The internal rate of return
on the equity element is projected as 13.7%-17.7% p.a. after tax. [Article by Theresa Ruhayel,
The Actuary, September 2014]
32. Residential property is a form of infrastructure which UK pension funds have tended to avoid
in the past, partly because of concerns about possible rent controls and partly because of the
reputation risks which might arise if the pension fund is perceived to be managing the properties
in ways which impact harshly on tenants. However, many of these risks could be mitigated by
using an intermediary management organisation. Many private individuals have found “buy to
let” to be a worthwhile investment and there seems to be no reason why pension funds and
insurance companies should not do likewise. In fact, as reported elsewhere in this paper,
insurance companies are already investing in student accommodation. The construction risks can
be largely avoided through fixed-price contracts. The rent control risk cannot be entirely
ignored, since such controls were proposed by the Labour Party prior to the 2015 general election.
Nevertheless it seems likely that residential property will become an increasing focus for
institutional investors, as they seek higher returns than can be obtained elsewhere. The gap
between available housing and the accommodation needs of the British population looks likely to
continue for many years, even if the rate of construction increases, so rents are unlikely to decline
and may even increase in real terms after allowing for inflation.
33. While UK financial institutions are likely to continue for a while to have only limited interest
in investing (other than through bonds) in forms of infrastructure new to them, as they gradually
acquire more experience their rate of investment in direct infrastructure ownership will probably
rise as they seek higher returns and diversification of risk in their overall investment portfolios.
They are likely to develop partnerships with other investors, and will seek to discover exciting
investment opportunities by forming close contacts with public officials who may be seeking
investment in projects they are developing. They will also develop partnerships with developers,
as the Malaysian pension fund did in the case of Battersea Power Station (see above). Another
partnership example is provided by Dutch pension fund asset manager APG, which in 2012
participated, alongside other parties, in a private equity fund focused exclusively on Philippine
infrastructure projects. It was announced that the fund will target 5 to 10 investments of between
$50 million and $125 million each, generating “attractive risk-adjusted returns” for pension funds
in the Netherlands.
19
34. One particular form of direct ownership is the holding of equity in a project where much of
the cost has been financed by an issue of debt. In this case the return to the investor will be
geared and the investor will need to consider the risks involved carefully.
35. Institutions which have investment tax privileges, such as UK pension funds, will need to
structure their direct ownership of infrastructure in such a way that they are not deemed to be
trading and hence liable to tax.
Listed infrastructure equities
36. Introduction
Listed infrastructure equity arguably provides an alternative method of gaining exposure to infrastructure
assets. Exposures can be gained either through listed infrastructure players (companies which own,
operate, manage or maintain physical infrastructure assets such as utilities, transportation, energy and
telecommunications companies), or through investment companies that invest in securities issued by
infrastructure Special Purpose Vehicles (SPVs).
Liquidity is one of the most attractive features because institutional investors can gain immediate access
to infrastructure, rather than waiting for years to strike a deal. A listed infrastructure portfolio is easily re-
adjusted according to the market conditions and institutional investors can divest at their own timing.
Exposure to infrastructure can be achieved at lower cost, and even offer better diversification across
different sectors and geographical regions, which cannot be obtained by small institutional investors in
unlisted infrastructure funds. Institutional investors will also benefit from daily pricing information and
detailed financial reporting of these listed companies.
However, it is very likely that infrastructure equities are already in the institutional investor’s existing
equity portfolio and the extent of those holdings would have to be taken into account in deciding on how
much extra should be invested in order to get the total exposure required. Listed infrastructure equities
can be impacted by equity market sentiments unrelated to infrastructure performance, which limit the
diversification benefit that infrastructure investment is meant to provide in the first place. Academic
literature has provided evidence of listed infrastructure equities being an inappropriate proxy for
infrastructure investment. Listed infrastructure equities exhibit high volatility, high correlation with
business cycles, low alpha, and limited downside and inflation protection – characteristics that do not
represent those of infrastructure investment.
37. Pure infrastructure players Pure infrastructure players consist of transportation, utilities, telecommunications and social companies
(as categorised by GICS classification or other) and also companies outside these sectors which are direct
beneficiaries of infrastructure activities (such as companies involved in shipping, building materials, power
generation etc.) .
Five major infrastructure indices are reviewed below and compared with an index of global equities as a
whole – MSCI ACWI Infrastructure, S&P Global Infrastructure REIT, Macquarie Global Infrastructure, UBS
World Infrastructure & Utilities, and Dow Jones Brookfield Global Infrastructure. All these indices use a
sector-based approach in filtering their constituents, except Dow Jones Brookfield Global Infrastructure,
which only considers companies that derive at least 70% of their cash flows from infrastructure activities.
It is important to note that these indices are market-cap weighted and are particularly dominated by US
20
0%
5%
10%
15%
20%
25%
30%
Dec
-07
May
-08
Oct
-08
Mar
-09
Au
g-0
9
Jan
-10
Jun
-10
No
v-1
0
Ap
r-1
1
Sep
-11
Feb
-12
Jul-
12
Dec
-12
May
-13
Oct
-13
Mar
-14
Au
g-1
4
Jan
-15
Rolling 5-year annualised volatility DJ Brookfield Global Infra. S&P Global REIT
FTSE All-World
companies and the utilities sector. Therefore, these indices may not represent the general infrastructure
market, but rather that of the utilities sector.
Investment performance
Figure1 compares the return and volatility of the Dow Jones Brookfield Global Infrastructure Index against the
S&P Global REIT index and the FTSE All-World index for data between January 2003 and April 2015. Dow Jones
Brookfield Global Infrastructure is chosen because its selective constituent process makes the index more
representative of the “infrastructure investment performance”. Throughout this period, the index has
demonstrated a higher rolling return and lower volatility as compared to the other two, although the excess
return has been narrowing since late 2013.
0%
5%
10%
15%
20%
25%
Dec
-07
May
-08
Oct
-08
Mar
-09
Au
g-0
9
Jan
-10
Jun
-10
No
v-10
Ap
r-11
Sep
-11
Feb
-12
Jul-
12
Dec
-12
May
-13
Oct
-13
Mar
-14
Au
g-1
4
Jan
-15
Rolling 5-year annualised volatility MSCI ACWI Infra. S&P Global Infra.
Macquarie Global Infra. UBS World Infra. & Util.
DJ Brookfield Global Infra. FTSE All-World
-15%
-10%
-5%
0%
5%
10%
15%
20%
25%
Dec
-07
May
-08
Oct
-08
Mar
-09
Au
g-0
9
Jan
-10
Jun
-10
No
v-1
0
Ap
r-1
1
Sep
-11
Feb
-12
Jul-
12
Dec
-12
May
-13
Oct
-13
Mar
-14
Au
g-1
4
Jan
-15
Rolling 5-year annualised return DJ Brookfield Global Infra. S&P Global REIT
FTSE All-World
Figure 1 (Source: Bloomberg)
Bloomberg)
-15%
-10%
-5%
0%
5%
10%
15%
Dec
-07
May
-08
Oct
-08
Mar
-09
Au
g-0
9
Jan
-10
Jun
-10
No
v-10
Ap
r-1
1
Sep
-11
Feb
-12
Jul-
12
Dec
-12
May
-13
Oct
-13
Mar
-14
Au
g-1
4
Jan
-15
Rolling 5-year annualised return in excess of FTSE All-World
MSCI ACWI Infra. S&P Global Infra.
Macquarie Global Infra. UBS World Infra. & Util.
DJ Brookfield Global Infra.
0%
20%
40%
60%
80%
100%
Dec
-07
May
-08
Oct
-08
Mar
-09
Au
g-0
9
Jan
-10
Jun
-10
No
v-10
Ap
r-11
Sep
-11
Feb
-12
Jul-
12
Dec
-12
May
-13
Oct
-13
Mar
-14
Au
g-1
4
Jan
-15
Rolling 5-year correlation against FTSE All-World
MSCI ACWI Infra. S&P Global Infra.
Macquarie Global Infra. UBS World Infra. & Util.
DJ Brookfield Global Infra.
21
Comparing all the major infrastructure indices against each other in the same period (in Figure 2), it
can be seen that only Dow Jones Brookfield Global Infrastructure consistently outperformed FTSE All-
World, while other indices registered negative excess return since 2010. These indices showed lower
volatility as compared to FTSE All-World, except perhaps for S&P Global Infrastructure. It is worth
noting that all five infrastructure indices actually displayed negative skewness and fat tails – the
skewness ranges from -0.88 and -1.12 while the kurtosis ranges from 1.5 to 2.8 in the same period.
Also the indices provide only limited diversification benefit, where their correlations against FTSE All-
World fall within 80%-100% range (albeit these are beginning to fall in recent months). These figures
suggest that the nature of these companies’ performance may not mimic the performance that we
would have expected from direct infrastructure investment.
We have also assessed the historic volatility using a Garch process (see Figure 3 below), to give an indication of the variation from month to month. This has been done for the five Infrastructure indices of Figure 2. For the three indices with higher overall volatility shown in red, dark blue and light blue, volatility spikes in 2008. This is especially true for the S&P Global Infrastructure index which reached a peak of 33%pa. For the two with lower overall volatility, not only was volatility lower but it was also flatter and more stable. The volatility of the MSCI ACWI Infrastructure index stayed within the low and narrow band from 10%pa to 16%pa. The fact that these two latter indices did not spike to the same extent as the others during the financial crisis of 2008 is perhaps somewhat surprising at first sight. (The assumptions underlying the fitted volatilities are a Garch model with Normal residuals.)
0%
10%
20%
30%
40%
Dec-2002 Sep-2005 Jun-2008 Mar-2011 Dec-2013
Vo
lati
lity
Date
Fitted Annualised Volatility Garch model with normal
residuals
MSCI ACWI Infra.
S&P Global Infra.
Figure 2 (Source: Bloomberg)
22
Figure 3 – Annualised volatility of infrastructure indices. Note: the colours are consistent with those shown in Figure 2 for the same indices.
In terms of dividend yield, infrastructure indices have a higher dividend yield as compared to global
equity and REIT indices. In January 2013 – April 2015 (April 2013 – April 2015 for Dow Jones
Brookfield Global Infrastructure), the 12-month dividend yield averaged 4.04% for S&P Global
Infrastructure, 4.09% for Macquarie Global Infrastructure and 3.28% for Dow Jones Brookfield Global
Infrastructure. For comparison, average 12-month dividend yields for the MSCI ACWI and S&P Global
REIT indices were 2.55% and 4.01% respectively. This is rather expected given the nature of low
growth stocks which would offer higher dividend yield.
The difference in the performance of infrastructure and equity indices during periods of high inflation
and low inflation are analysed in Figure 4 below . Inflation is measured by monthly changes in the
Consumer Price Index of the respective region, from January 2004 to March 2015. Months where the
monthly inflation is above the median inflation are categorised as “high inflation” and vice versa. The
average for the monthly returns of the infrastructure index is compared to that of the equity index for
each inflation period. The charts below show that in high-inflation periods, the infrastructure index
outperformed the equity index, although this was more significant in the US than in Europe. (Note:
Macquarie Global Infrastructure North America is chosen to represent the US market because of its
longer available history. US constituents are dominant in the index, so the index should fairly represent
the performance of US infrastructure companies.) Numerous (and more robust) attempts have been
made to assess the inflation protection that listed infrastructure equities could provide and this is
merely a simple analysis to observe how the returns behave in different inflation periods.
0.0%
0.5%
1.0%
1.5%
High inflation Low inflation
Average monthly returns of infra. index and equity index in high inflation and low
inflation periods in the US
Macquarie Global Infra North America S&P US Composite 1500
0.0%
0.5%
1.0%
1.5%
High inflation Low inflation
Average monthly returns of infra. index and equity index in high inflation and low
inflation periods in Europe
Macquarie Global Infra Europe FTSE Europe
Figure 4 (Source: Bloomberg)
23
38. Investment companies
These companies are closed-end investment companies which invest in equities and/or subordinated debts issued by infrastructure SPVs. The proceeds from the equity and debt issuance are used to finance the construction and/or operation of infrastructure projects commissioned by the public sector.
The advantages of accessing the asset class through an investment manager are: 1) managers are normally specialized and have access to resources and expertise otherwise unavailable to other investors; 2) as investment managers pool funds, investors can get access to both larger and smaller projects, and benefit from a diversified portfolio. These advantages are particularly attractive to smaller institutional investors, which may lack resources in having their own in-house investment specialist to carry out the necessary research analysis. However, investors in pooled funds may have limited control over the choice of the underlying infrastructure exposure.
Closed-end investment companies listed on the London Stock Exchange are:
Company Market focus Geographic focus Market cap1
(GBP mn)
3i Infrastructure (3IN) Core economic infrastructure (utilities and transportation sectors), social infrastructure, primary PPP
Europe and India 1,390
Bilfinger Berger Global Infrastructure (BBGI)
PPP/PFI projects Europe, North America, Australia and New Zealand
530
HICL Infrastructure (HICL)
Operational social and transportation infrastructure (PPP/PFI)
UK, Europe, North America and Australasia
2,010
International Public Partnerships (INPP)
Social infrastructure UK, Europe, North America and Australasia
1,160
John Laing Infrastructure (JLIF)
Equity and subordinated debt of PPP projects
UK, Continental Europe and North America
1,030
GCP Infrastructure (GCP)
Infrastructure debt UK 609
Sequoia Economic Infrastructure Income (SEQI)
Senior and subordinated debt of economic infrastructure
Europe, US, Australia and New Zealand
159
Vietnam Infrastructure (VNI)
Shares of companies involving in infrastructure, or infrastructure funds
Vietnam, China, Laos and Cambodia
172
Source: Company websites
1 As of March 19th, 2015
24
Fees, dividend policy and hurdle rates are as follows:
Company Ongoing charge ratio
Dividend policy targets Investment projects’ hurdle rates
3i 1.45% 5.5% yield IRR 7-8% net of expenses
BBGI 0.98% 5.5% yield 7 – 8%
HICL 1.12% 4.55% yield 7 – 8%
INPP 1.29% Long-term inflation expectation (2.5%)
8 – 9%
JLIF 1.35% 6% yield on IPO issue price Weighted average rate, based on government bonds. This was 8.19% for 2014, based on 3.46% risk-free rate + 4.73% risk premium.
GCP Infrastructure
N/A Up to 8 pence per share (e.g. 6.79% as of April 24
th, 2015)
N/A. Minimum reported yield on the investment portfolio is 9%
Sequoia Economic Infrastructure Income
N/A 5% in the first year of operation (2015) and 6% thereafter
N/A. Typical return quoted ranges from 6 to 9% as of February 2015
Vietnam Infrastructure
2.6% N/A. 2014 dividend yield was 5.6%
N/A
Source: factsheets and annual reports 2014
Discount rates used to value the investment companies’ projects reflect these factors:
The difference between the discount rate and the risk-free rate gives the implied risk premium of the projects.
The risk premium reflects the concession life of the projects (the longer concession life i.e. less mature projects, the higher risk premium), construction risk (if the project is still at the construction stage), inflation sensitivity, and the level of gearing within the SPV of the infrastructure projects
Different ranges of discount rates for different types of infrastructure assets: operational assets, assets going into operation and assets under construction, across all investment companies
The following figures for the volatility and return of five of these investment companies are interesting but
of limited use because they relate only to relatively short periods of time and cannot therefore be taken as
representative of what might be expected over longer periods at different phases of the economic cycle.
Company 3i BBGI HICL INPP JLIF
Period Jan’10 – Dec’14
Jan’12 – Dec’14
Jan’10 – Dec’14
Jan’10 – Dec’14
Jan’11 – Dec’14
Total return / excess return over FTSE all-share (%)
2
14.1 / 5 11.6 / 0.7 12 / 2.9 8.6 / -0.5 9.3 / - 2.2
Annualised price volatility / excess over FTSE all-share volatility (%)
8.3 / -4.1 7.6 / -2.8 6.8 / -5.6 6.1 / -6.3 6.4 / -4.8
Return per unit of volatility / excess over FTSE all-share
1.7 / 0.97 1.52 / 0.48 1.76 / 1.03 1.41 / 0.67 1.45 / 0.42
Correlation with FTSE All Share 0.10 0.30 -0.03 0.15 0.14
2 Total return includes price and dividend return. All dividends are reinvested in stock.
25
39. Summary
The above analysis suggests that listed infrastructure equities have liquidity advantages over
direct investment in infrastructure, but they do not offer much in the way of diversification or in
most cases outperformance, compared with equities as a whole, though there is some evidence
that they have lower volatility.
40. References:
1. Anagnos, J. and De Croisset, Marc. (May 2014). Global listed infrastructure: The case for an allocation
by institutional investors. CBRE Clarion Securities
2. Blanc-Brude, F. (January 2013). Towards efficient benchmarks for infrastructure equity investments.
EDHEC Risk Institute Publication (pg 36–57)
3. Blanc-Brude, F. (May 2014). Making a better match between institutional investors and infrastructure
investments. Rethinking infrastructure: Voices from the global infrastructure initiative. McKinsey &
Company (pg 15 – 22)
4. Cheng, T. and Srivastava, V. (April 2015). Approaches to benchmarking listed infrastructure. S&P Dow
Jones Indices
Overall portfolio construction
41. Once an institution has taken a decision in principle to invest in infrastructure, the question
will arise as to how much to commit to this class, as well as the split between bonds and equity.
Amendments to the institution’s risk appetite statement may be needed. If the decision is to
invest only in infrastructure bonds, the bonds will be little different from those of other
organisations with the same credit rating, though they may differ in terms of cash-flow profile and
recovery rate. However, unless there is a full Government guarantee of the bonds, consideration
will have to be given to any big uncertainties to which the infrastructure itself is exposed (see
paragraphs 12-22) and which may not have been given as much weight, in the relevant rating
criteria used to assess the credit quality of the asset and generate the credit rating, as the investor
might attribute to them. Such uncertainties, however remote they may appear today, may cause
an investor to hesitate before committing too high a proportion of its portfolio to infrastructure
bonds, even if they offer a marginally higher return than Government bonds. In cases where the
credit rating is less than prime, the proportion of the portfolio to be invested in such bonds will
have to be considered in the same way as corporate bonds. It will need to be remembered that
any apparent liquidity might dry up if the underlying infrastructure asset is itself seen to be in
trouble.
42. Different considerations arise in the case of direct ownership of infrastructure assets, either
freehold or leasehold, where the risks and potential returns will be much higher, particularly if
there are construction and forecasting risks. The investor will need to make a judgement about
the risk-reward balance on each individual asset and whether the investment is likely to be
worthwhile. A good spread of assets to achieve diversification of risk between different assets is
desirable, and in the case of a smaller investor this can probably best be achieved by investing
26
through a platform in partnership with other investors, sharing the cost of due diligence with
them. Only the largest institutions are likely to be able to afford to build up a diverse portfolio of
infrastructure investments on their own. In those cases where institutions already invest in
commercial properties, it may make sense to reclassify them as infrastructure and consider them
as being in the same category as the new forms of infrastructure now being introduced to the
portfolio.
43. One of the considerations which usually has to be taken into account when deciding on the
allocations between different classes of investment is the extent to which the returns of the
various classes will be correlated with each other. This is particularly important in those cases
where an annual performance report is made to members or whether the assets are valued at
market value for consolidation into the accounts of a sponsoring company. Just like commercial
property, any directly-owned infrastructure assets will need to be revalued from time to time for
such purposes, but capital values are likely to lag behind equities when values move upwards or
downwards with the economic cycle. This stability may be more apparent than real, but
nevertheless could be seen as an attraction of infrastructure.
44. Another decision which will need to be made when going into direct ownership is whether
there are any preferences between different classes of infrastructure, for example are energy
investments to be preferred over transport, student accommodation or “buy to let”, or is a good
spread over all classes desired to achieve greater diversification of risk? Is the investment to be
restricted to a long lease of infrastructure which has already been developed or is it to extend to
green field sites on which the assets are to be constructed? Will overseas assets be considered?
45. In deciding on the prospective return which would be acceptable from direct ownership of
infrastructure, we suggest that where the risks are considered to be broadly similar to those of
commercial property, the investor might well be looking for a similar return. This might be
enhanced in cases where there is less liquidity. At the end of 2014 the running yields on UK
prime commercial properties averaged 5.6% per annum [Source: CBRE Market View, UK Prime
Rent and Yield, Q4 2014]. The yields on good secondary properties would have been higher
than this, though with greater risk. An analysis of the prospects for long-lease property by M&G
Investments [Pensions Age, April 2015, page 63] over the period 2014-2048 indicated that the
return to an investor would be 4.8% per annum, even if the investment had to be written off at the
end of the period. If the property value remained unchanged, the return would be 6.9% per
annum. It should be borne in mind that these figures might be somewhat enhanced in due
course, due to rent increases.
46. Using the returns on commercial property as a benchmark, we can start to make an estimate
of the kind of return we might want to get from a long lease on an infrastructure investment which
is thought to bear a similar degree of risk. Perhaps the purchaser of the lease of a UK
infrastructure asset which has already been constructed and has a good-quality income stream
might therefore want to get an initial yield of around 6 per cent per annum, assuming that the
asset reverts to the public sector at the end of 30 years with no return of capital to the investor. If
the income stream is less certain than a prime property rent, or if the asset has yet to be
constructed, the prospective yield would have to be higher, to allow for the substantial extra risks
involved. If the income stream is linked to inflation, a slightly lower initial yield might be
contemplated.
27
47. In the case of infrastructure bonds with a strong public sector guarantee, the return required
will be significantly lower, reflecting the lower risk and the eventual capital repayment.
However, as indicated earlier, there will still usually be a degree of risk and lower liquidity than
in the case of government bonds, necessitating a higher rate of return than on those bonds, to
compensate.
48. In the case of equity infrastructure investments as part of the capital structure of special
purpose vehicles, the projected rates of return will have to be treated with considerable caution,
given all the risks involved. However, it may be that a carefully selected portfolio of such
investments, chosen to spread the risks as far as possible and with plenty of upside potential for
extra development, will give good returns in the long run. At the very least it will introduce a
degree of diversification into the overall portfolio.
Practical issues
49. There are many practical issues for an investor in infrastructure, including the need to find
suitable investments with the right durations, identify the risks, exercise due diligence, monitor
the investment as time goes on, etc. Thought will have to be given to the choice of route:
bonds, listed companies, infrastructure funds, direct ownership, or participation in equity. A
policy will have to be established on the classes and locations of the infrastructure which will be
considered. The work involved in direct ownership will often be beyond the capabilities of an
individual investor, and a partnership with others may be necessary, perhaps via a platform. The
investor’s own regulatory issues will need to be considered. As part of the due diligence process
for direct ownership, it might be worthwhile to build a model, showing the income streams which
are expected and then testing the model to see what might happen in various future scenarios. In
cases where the asset has yet to be constructed, the investor may wish to insist on the use of a risk
management methodology such as RAMP.
50. One of the most difficult areas at present is the finding of suitable investment opportunities.
In the case of infrastructure bonds there is such strong demand that in some cases yields are little
higher than those on gilts, and it is perhaps questionable whether they provide an adequate return
to the investor. In the case of direct ownership, some institutions have started to develop
relationships with public officials, so that they can hear of projects at an early stage. This also
helps the officials, since they will have a better idea, before they have done a great deal of work,
on whether their project is likely to attract investment. Such liaison between officials and
investors is likely to continue if the UK continues the process of devolution in ways which enable
local authorities to develop and seek funding for their own projects without the necessity to seek
Government approval for each one. This could lead to a multitude of smaller projects becoming
available for investment. In cases where the asset has yet to be constructed, we may see
partnerships developing between developers, who will bear the construction and forecasting risks,
and long-term investors who will buy out the completed asset.
51. Once the investment has been made, there may be a need to continue to keep a careful watch
on it, and this will take suitably experienced resources. Sometimes there may be a need to vary
the terms initially agreed, for example.
28
Possible future work
52. The working party has not yet decided on whether it wishes to study other aspects of
infrastructure investment, and would welcome views from the readers of this paper. Some of the
areas which could be considered for research are:
questions of sustainability, having regard to climate change and other major uncertainties,
both from the perspective of risk as well as the opportunity and duty to be constructive
in helping to enable the construction of resilient and adaptable infrastructure;
a study of whether particular avenues of direct ownership are likely to be especially
attractive for investors;
a study of the extent to which institutions might be able to participate in infrastructure
investment in underdeveloped countries, having regard to problems such as corruption,
lack of financial structures, extreme political risk, etc. Are there structures which could
be developed that would enable some of these risks to be minimised and enable
institutions to invest?
Further reading
53. We recommend a useful paper prepared recently by an IFoA working party headed by Gareth Mee: Non-traditional investments – key considerations for insurers, January 2015.
[available on-line at http://www.actuaries.org.uk/research-and-resources/documents/non-
traditional-investments-key-considerations-insurers-long-versio]. Although the paper deals
with a variety of non-traditional investments, there are some excellent sections on
infrastructure, including links to online case studies.
The RAMP methodology referred to in paragraphs 20 and 49 is set out in a guide published
by ICE Publishing, Institution of Civil Engineers, 3rd
edition, 2014. It was developed by a
working party of actuaries and civil engineers, and is used by London’s Crossrail as a basis
for its own risk-management system.
54. Here are some useful web-sites:
http://www.oecd.org/pensions/private-pensions/institutionalinvestorsandlong-terminvestment.htm
http://www.worldbank.org/en/news/press-release/2014/10/09/world-bank-group-launches-new-
global-infrastructure-facility
http://www.oliverwyman.com/insights/publications/2012/apr/pension-funds-investment-in-
infrastructure.html#.VSfOAfDE12s
http://www.weforum.org/reports/infrastructure-investment-policy-blueprint
http://www.iiasa.ac.at/web/home/research/researchPrograms/RiskPolicyandVulnerability/Flood_Re
silience.en.html
29
http://www.oliverwyman.com/insights/publications/2012/mar/strategic-transport-infrastructure-
needs-to-2030.html#.VSfPRvDE12s
http://www.oliverwyman.com/insights/publications/2014/mar/oecd-institutional-investors-and-
infrastructure-financing.html#.VSfPZvDE12s
http://www.oliverwyman.com/insights/publications/2014/jan/infrastructure-development-
davos.html#.VSfPjvDE12s
http://www.nerc.ac.uk/innovation/activities/infrastructure/envrisks/
Conclusion
55. It is clear that the main obstacle which is preventing institutions from investing more
money in infrastructure in developed countries at present is a shortage of suitable
investments. Yields on infrastructure bonds are currently so low that some may question
whether the return is high enough to justify investing in them, though they may still be
useful for insurance companies and pension schemes looking to match long-term liabilities.
There is also a lack of willingness to invest in any form of infrastructure on the part of some
institutions, because they do not have an adequate understanding of the risks involved, and
hopefully this paper may start to fill in some of the information they need. Much of the
infrastructure investment which the world requires over the next 15-20 years will be in
countries which are still developing, and it is here that some of the biggest returns may be
achievable for those investors who are prepared to take on the formidable risks involved
and make selective investments after careful due diligence
Working Party members:
Chris Lewin (chairman), Dalila Hashim, Clara Hughes, Malachy Magennis, Bruce Porteous,
Theresa Ruhayel, Francisco Sebastian, Cliff Speed, Brendan Walsh, Joshua Waters, Clara
Yan, Tracey Zalk.